The Eye of the Master
In The You On AI Field Guide
Pasquinelli credits Federici among the feminist scholars who 'have explained the rise of modern rationality and mechanical thinking (to which AI also belongs) in relation to the rule of women's bodies and the transformation of the collective body into a docile and productive machine.' The connection is direct: the same historical process that disciplined bodies for labor also developed the techniques for extracting knowledge from bodies and encoding it in machines. The Taylorist decomposition of skilled work into measurable operations, the Fordist assembly line that replaced craft knowledge with mechanical routine, the computerization that automated clerical work — each represented a phase in the ongoing project of separating intelligence from the human beings who possessed it.
The book's title references Bentham's panopticon and its principle that observation is a technology of power. The master's eye — the surveillant gaze that records, measures, and controls — is the instrument through which capital has always learned. Workers under observation perform their knowledge visibly. The observation captures the knowledge in records, specifications, time-motion studies, and now training data. The captured knowledge becomes the basis for systems that can perform the work without the worker. The worker is displaced, and the displacement is achieved through the worker's own labor: the knowledge that made the worker valuable becomes the instrument of the worker's replacement.
Applied to contemporary AI, Pasquinelli's framework reveals that the training of language models on the collective output of programmers, writers, and knowledge workers is not a novel development but the perfection of a technique capital has been refining for two hundred years. The GitHub repository is the panopticon. The public internet is the commons being observed. And the large language models trained on this material are the crystallization of collective human intelligence into private capital that will be deployed to reduce capital's dependence on the human labor that produced it.
The genealogy is political, not merely historical. By demonstrating that AI is continuous with previous waves of knowledge extraction and worker displacement, Pasquinelli refuses the narrative of technological inevitability. The dispossession is not the unintended consequence of neutral tools. It is the continuation of capital's deliberate, centuries-long project of appropriating collective knowledge. The project can be recognized, resisted, and restructured — but only if its historical continuity is understood and only if the resistance is organized at the scale of the project itself.
Origin
Pasquinelli is an Italian philosopher based in Germany, affiliated with the Karlsruhe University of Arts and Design. The Eye of the Master was published by Verso in 2023 and rapidly became one of the most cited works in critical AI studies. The book synthesizes labor history, history of technology, and Marxist political economy into a genealogy demonstrating that artificial intelligence is not an autonomous development but a chapter in capital's continuous effort to extract and mechanize human knowledge. Federici endorsed the work as 'a crucial contribution' to understanding AI's political economy.
Key Ideas
AI is the latest phase of knowledge extraction. From Babbage to neural networks, capital has pursued the systematic recording of workers' knowledge and its encoding in machines that replace those workers.
The master's eye is the instrument. Surveillance, measurement, and observation are not neutral but political — the mechanisms through which capital learns from labor in order to dispossess labor.
Workers teach the machine that replaces them. The training data — collective code, writing, images — represents millions of workers' knowledge, appropriated to build systems that will compete with those workers.
Dispossession is epistemological. Workers lose not only jobs but sovereignty over their collective intelligence, which is absorbed into proprietary systems they do not own or control.